Two-Layer Fault Diagnosis for Storage Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Storage systems face inefficiencies in fault diagnosis, as faults caused by environmental factors often require substantial time and resources from customer support teams, and existing technologies lack the ability for the systems to quickly pre-diagnose and resolve these issues locally.
Innovation Solution
A 2-layer fault diagnosis system is implemented, with a local pre-diagnosis model deployed in the storage system and a cloud-based diagnosis model, where the local model is distilled from the cloud model, allowing for quick identification and local resolution of faults caused by environmental factors, thereby reducing the workload on customer support teams.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a cloud-based diagnosis system is used for all faults, then diagnosis accuracy is improved, but response time and support team workload increase
Solution Approach 1:
The diagnosis system is segmented into two parts: a lightweight first diagnosis model deployed locally in the storage system for immediate fault screening, and a comprehensive second diagnosis model deployed in the cloud for detailed analysis. This segmentation allows common faults to be diagnosed locally without cloud dependency, reducing response time while maintaining accuracy for complex cases through the cloud-based model.
Solution Approach 2:
The first diagnosis model performs preliminary diagnosis of faults locally before they reach the cloud-based system. By pre-screening faults and identifying those that can be resolved locally, the system performs useful action in advance, reducing the burden on the cloud-based diagnosis system and accelerating response time for common issues.
2Loss of time
If a lightweight local diagnosis model is deployed, then response time is improved, but diagnosis capability is reduced
Solution Approach 1:
The first diagnosis model acts as an intermediary between the storage system and the cloud-based second diagnosis model. It screens faults locally and selectively transfers only those requiring advanced diagnosis to the cloud, maintaining a balance between fast local response and comprehensive cloud-based analysis capability.
Solution Approach 2:
The system adds a spatial dimension to diagnosis by deploying models at two different locations (local storage system and cloud). The first diagnosis model handles immediate local decisions, while the second diagnosis model provides comprehensive cloud-based analysis, creating a multi-dimensional diagnosis architecture that balances speed and capability.
3Reliability
If all faults are handled by customer support teams, then comprehensive diagnosis is ensured, but resource consumption increases
Solution Approach 1:
The storage system performs self-diagnosis using the locally deployed first diagnosis model, which can identify and resolve common faults caused by environmental factors without human intervention. This self-service capability reduces the workload on customer support teams while maintaining comprehensive diagnosis through the cloud-based second model for complex cases.
Solution Approach 2:
The first diagnosis model discards (filters out) common faults that can be resolved locally, preventing them from reaching the customer support team. By selectively discarding easily resolvable issues, the system reduces support team resource consumption while maintaining the ability to handle complex cases through the cloud-based model.
Data Source
AI summary
A method in an illustrative embodiment of the present disclosure includes determining, utilizing a first diagnosis model deployed in a storage system, whether a cause of a fault belongs to environmental factors. The method further includes determining, responsive to determining that the cause of the fault belongs to the environmental factors, whether the fault can be solved locally in the storage system. The method further includes sending, responsive to determining that the fault cannot be solved locally in the storage system, the fault to a second diagnosis model, wherein the first diagnosis model is obtained by distilling the second diagnosis model. According to the method for fault diagnosis of the present disclosure, particular faults can be diagnosed and solved locally in a storage system, so that the workload of a customer support team of the storage system in a cloud can be reduced.